The literature recognizes several main groups of religious traditions: Catholics (Herberg 1955; Turner 1971; Levine 1986); Mainline Protestants (Weber [1905/1958]; Herberg 1955; Layman 1997); Evangelicals (Martin 1990; Stoll 1993: Campbell 2006); and non‐Christian (Herberg 1955; Lenski 1963). Unfortunately, those classifications do not offer clear answers about how to classify specific denominations.  For example, it does not clearly indicate whether AmericasBarometer Insights: 2009 (No. 29) the American Baptist Church is part of the tradition (Layman 1997) or Methodology Note: part of the Evangelical tradition (Campbell 2006). In the same vein, it is not always clear Measuring how to categorize Latin American Baptist Churches, and whether or not Latter‐Day Saints should be considered as part of Protestant or in Surveys of the Evangelical traditions.

Americas If we take religion to be a multifaceted phenomenon that includes believing, behaving,

Alejandro Díaz‐Domínguez and belonging (Layman 2001), then, Vanderbilt University measurements of religious affiliation capture the alejandro.diaz‐[email protected] ”belonging” component, which refers to an individual’s religious membership and group he ability to distinguish among survey identification. This paper argues that it is respondents’ religious affiliations is a key important to classify religious affiliations with Tprerequisite to studies of the effects of as much precision as possible. Further, the religious traditions on economic development, paper indicates how these classifications can be political preferences, and support for related to economics and politics. democracy. While the impact of religion on several spheres of human life has long been Theoretical Background studied, the measurement of religious affiliation at the individual level remains problematical. Considering the low number of articles published about religion and politics in the top This paper presents some theoretical and political science journals over the past 30 years, empirical perspectives that have led to a new one might be tempted to conclude that the topic AmericasBarometer questionnaire item for is of little significance. However, some argue classification of religious affiliations in the that a primary reason for this low number of region. This classification will be used for the publications is that specialists on religion have first time in the AmericasBarometer 2010. often failed to link their work to broader theories of political behavior and often do not work with data of the best possible quality (Wald and Wilcox 2006: 529). This paper  addresses the latter problem related to data by The Insights Series is co‐edited by Professors Mitchell A. Seligson and Elizabeth Zechmeister with administrative, discussing the importance of classifying technical, and intellectual support from the LAPOP group at religious denominations with greater precision, Vanderbilt University. © 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 1 of 13 and the former theoretical situation by sketching arguably illustrative and important to identify the connection between religious traditions and the largest religion for political reasons. economics and politics. In modern political analysis, religion has also Beginning with economics, there is a been related to several political processes and longstanding view that considers as useful to outcomes. These include Catholics’ political role distinguish between Catholics and Protestant in Latin American democratic transitions denominations when evaluating the impact of (Huntington 1989); competition in the religious religion on economic accumulation because market among Evangelical and Catholic those religious differences underlie different Churches, where Catholic leaders pushed for capacities for stimulating economic growth, as democracy instead of authoritarianism in order Weber [1905(1958)] argues in his classic work, to retain parishioners in (Gill Protestant Ethic and the Spirit of Capitalism. 1998 and 2001; Lies 2006); and, finally, the collaborative pro‐democratic efforts of Catholic Several social scientists have attempted to clergy and congregants (Turner 1971; identify the impact of religion on economic Mainwairing 1986; Camp 1997; Hagopian 2008). growth, with varying degrees of success. Findings from these studies include, for Another perspective on politics, economics, and example, mixed effects of Catholicism on Catholicism is presented by Levine (1986), in his poverty (Banfield 1958: 87 and 123; Przeworski Religion and Political Conflict in Latin America. et al 1998; Barro and McCleary 2003); a positive Levine argues that rank‐and‐file Catholics, who relevance of for economic growth are generally poor, demanded recognition from (Inglehart 1997; Harrison 2000: 99); a the Catholic hierarchy through class‐based relationship between successful economic organizations, and Catholic Bishops therefore policies and Confucianism (Swank 1996; Kahn created democratic spaces inside and outside the 1979); tracing the origins of Japanese capitalism Church. In sum, links between democratization, to Buddhist foundations (Collins 1997); linking a class, and religious affiliation suggest the tradition of literacy among of Eastern necessity of capturing Catholicism not only Europe to how they overcame a “subculture of because it represents the majoritarian religion in poverty” (Lewis 1966: 24); and, as a several Latin American countries, but also potential vehicle of economic growth (Nolan because Catholicism plays a relevant role in the 2005). political arena.

With respect to the impact of religion on politics, In two other works, Protestant, Catholic, Jew conventional wisdom at times points to Marx’s (Herberg 1955) and The Religious Factor (Lenski well‐known statement defining religion as the 1963), both of which study religion in the US, opiate of the people [1843(1978)]. However, the critical differences were shown among these political relevance of the largest religious three main affiliations. At the same time, in affiliation was initially discussed in Marx’s On Latin America, Tongues of Fire (Martin 1990) and the Jewish Question, where he argues that the incorporation of the study of religious emancipation secularizes political (see also Stoll 1993) have further motivated the relationships (this emancipation implies the debate about how to correctly classify religious equal co‐existence of several religions and denominations, especially because of their precludes the largest religion from becoming a various theorized political implications.1 Given political majority). It is relevant to note that Marx does not detail the specific role of any 1 For example, it may be relevant to distinguish among particular denomination. Nonetheless, it is diverse types of Pentecostals and Neo‐Pentecostals in terms of support for authoritarian rule, as in the case of Guatemala © 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 2 of 13 both the recognition that different religious it is to obtain as much detailed data as possible, orientations and communities may differently and do so as parsimoniously as we can. affect economic and political processes, as well as the recent increase in the number of religious The research for this paper was carried out in dominations present in Latin America, it is of three stages. The first stage was to formulate a real importance that data exist to capture these comprehensive list of religious denominations affiliations and memberships. and potential variants of labels identifying non‐ religious persons in Latin America. The second In sum, placing the study of religion and politics stage was to identify the proportion of affiliates or economics in a theoretical perspective is an by denomination using a minimum threshold of important step for understanding the different 2% of affiliates within a given national implications of the religious diversity in terms of population (Layman 1997). The final stage was the sense of religious belonging that diverse to classify or categorize religious traditions for religious traditions represent. use in the AmericasBarometer 2010 survey.

The Empirical Challenge: The natural starting point in terms of sources is Capturing emerging religious the national census, and I consulted all those instruments when the information was groups available.3 Nevertheless, some national censuses do not include the question: “What is your Correctly classifying religious affiliations at the religion?”4 individual level presents two main challenges: a) to include the most common religious The second best source of information on denominations; and b) to categorize correctly religious affiliation are national surveys each religious denomination. The designed to capture some degree of religious AmericasBarometer surveys2, carried out by the diversity, such as the 2008 round of the Latin American Public Opinion Project AmericasBarometer (Cruz 2009); the Central (LAPOP), have always striven toward these dual American Public Opinion Project (Stein 2000);5 goals but a modified set of questions will the World Values Survey (WVS) information in increase the extent to which we can meet them. some Latin American samples (Gill 2002; Magaloni and Moreno 2003);6 some Given that there are many denominations, LAPOP aims to include in its survey question on 3 The most detailed examples were Bolivia (the 2001 and religion as many religious beliefs as possible 2002 Censuses); (2000 Religiao Census); (2000 within and across traditions, nations, and Census); and Nicaragua (2005 Census). However, as I will mention later, usually the aggregation of religious traditions cultures. However, since classifying religion is prevents us from knowing the proportion of affiliates for only one of a wide variety of pieces of specific denominations. information that are gather in these hour‐long 4 Some examples are the national censuses of Colombia, surveys, trade‐offs are necessary. Our goal, then, Venezuela, and Argentina. 5 In the case of , the surveys consulted were the 1991, 1995 and 1999 Salvadorian samples; the 1992, 1993, 1995 and 1997 Guatemalan samples; the 1991 Honduran (Hallum 2002). However, this distinction is not always sample; and, the 1991 and 1995 Nicaraguan samples. possible in practical terms because of the small numbers of 6 In particular, the V144 question in the WVS questionnaire, cases (Sherman 1997) in any sample. using the 1995 and 1999 Argentinean samples; the 1997 2 Funding for the AmericasBarometer project is mainly Brazilian sample; the 1996 and 2000 Chilean samples; the provided by the United States Agency for International 1997 and 1998 Colombian samples; the 1996 Dominican Development (USAID). Other important sources of support sample; the 1999 Salvadorian sample; the 1990, 1996, 2000 are the Inter‐American Development Bank (IADB), the and 2005 Mexican samples; the 1996 and 2001 Peruvian United Nations Development Programme (UNDP), and samples; the 1996 Uruguayan sample; and finally, the 1996 Vanderbilt University. and 2000 Venezuelan samples. © 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 3 of 13 Latinobarómetro’ surveys;7 and specialized Affiliations in the surveys at the national level.8 In order to show the state of the religious The second task is to identify religious diversity in the Americas, as it is currently denominations that have at least a minimum known, we present results from the number of adherents, a level we set at 2%. This AmericasBarometer 2008 survey in Table 1 (see was possible using census data where this the appendix), sorted by religious category and information was available9 and survey data for alphabetically by country12. Those results show Catholicism in almost all countries; specific the importance of classifying with more detail Pentecostal Churches in Brazil, Guatemala, and native or traditional religions, which turn out to Chile; and the proportion of secularists or non‐ represent more than 2% of affiliates in Belize, believers in nearly all of Latin America. Brazil, Haiti, and Jamaica, given that they reach our theoretical threshold of 2% of affiliates In sum, identifying the proportion of religious within a given national population. affiliates at the national level was possible using census data10 and surveys.11 Nevertheless, a There is also room for increased precision in the common denominator in those sources is that LDS/Jehovah’s Witnesses category, for which the religious information is grouped in very few Belize, Chile, Jamaica, and Peru exceeded the categories. This fact also speaks to the threshold of 2% as Table 1 shows. LDS and fundamental need for data that classify religious Jehovah’s Witness are best considered as denominations with greater precision, as this different types of religions (Gill 2002), and project proposes. therefore it is important to create a special category for each.13

7 Specifically, I used the S2 open question in the With respect to and eastern religions, Latinobarómetro (LB) questionnaire, and the list of affiliations contained in the variable S42 since 1995 until the AmericasBarometer practice has been 2007. grouped them as non‐Christian Eastern religions 8 Specialized surveys consulted by country: 2008 because of the limited number of cases,14 but CEIL/CONICET survey about beliefs and religious attitudes theoretically speaking Judaism is in Argentina; 2004 IBOPE survey about Creationism in Brazil; 2006 Pew Center survey of Pentecostals in Brazil, distinguishable from Eastern religions (Nolan Chile, and Guatemala; and finally, 2003 Parametría Omnibus 2005). For that reason, Judaism will be survey, and 2007 CONSULTA survey about religious considered as a different category. practices in Mexico. In addition, Guatemalan and Costa Rican specific Protestant traditions were extracted from PROLADES (www.prolades.com). Non‐religious labels were As noted above, among Catholics it is possible obtained from the 1990 Church‐State Relations survey to distinguish among several types, such as conducted in Mexico (Camp 1997). Finally, fundamentals “true Catholics” and those who are not about traditional religions such as voodoo were collected observant. However, in order to retain the from Fontus (2001). 9 For example, one useful distinction among Protestant largest religion in Latin America as a single affiliations based on census data was the Moravian Church category for comparative reasons, this in Nicaragua with 1.63% of affiliates. distinction should be made by means of another 10 The aggregation is practically the rule and not the exception in surveys and Censuses. Perhaps, one remarkable exception is Brazil with its list of Evangelical churches, 12 The specific non‐response to Q3 question in the 2008 whereas Bolivia and Mexico Censuses are a good example of round was 0.9% for 21 Latin American countries. aggregation. In the last two cases, only the names of 13 My theoretical distinction does not follow the label created religions considered in each group was available. by Bastian (1993) about LDS and Jehovah’s Witness (“Para‐ 11 Essentially, surveys aggregate affiliations because the Christians”) given its negative connotation. reduced number of cases (Gill 2002). However, the name of 14 According to the World Christian Database, in 21 Latin several religious denominations included in some categories American countries covered by the AmericasBarometer, was available. Judaism only accounts for 430,000 affiliates. © 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 4 of 13 variable, i.e. the church attendance variable (Q5a variable at the country level, and apply an in the AmericasBarometer questionnaire). analysis using a Hierarchical Linear Model. In the next five figures, I will show the effects The main controversy among religious studies is obtained from adding additional religious how to classify Protestant or Evangelical categories to predict support for democracy. denominations. Here we will follow Weber’s original conceptualization (1905/1958) and The significance of the variables in the model is Layman’s classification for the US case (1997); graphically represented in the next five figures. this classification also took into account Gill’s Statistical significance is captured by a concerns about few cases (2002); and Sherman’s confidence interval that does not overlap the (1997) and Hallum’s (2002) debates about vertical “0” line (at .05 or better). When the dot, Pentecostals.15 which represents the predicted impact of that variable, falls to the right of the vertical “0” line, The Current State of the Art it implies a positive relationship whereas when it falls to the left, it indicates a negative As mentioned at the beginning of this paper, contribution. measuring religious affiliation captures the ”belonging” component, which represents Figure 1 shows that being Catholic decreases individuals’ sense of membership and group support for democracy. This analysis also identification. In order to justify empirically suggests that being non‐Catholic increases why it is important to classify religious support for democracy. However, with this affiliations correctly, I analyze six main religious distinction alone, we do not know what types of categories in order to see whether or not those non‐Catholic citizens are more likely to support religious categories are related to politics, in democracy. particular, to support for democracy.

The conceptual operationalization of support for Figure 1. democracy in terms of political culture used by Support for Democracy in the Americas, 2008 (Catholics and Non‐Catholics) the AmericasBarometer explicitly measures whether or not Latin American citizens believe Support for Democracy 2008 that democracy is better than any alternative Catholic form of government based on the question Church Attendance Rural to Urban Wealth Index wording developed by Mishler and Rose (1999; Female Percep Nat Eco Situation see also Rose and Shin 2001). The specific Percep of Security Political Interest question wording is “Democracy may have Education Level Interpersonal Trust problems, but it is better than any other form of Political Knowledge Index Trust Cath Church government. On a scale from 1 to 7, to what Age Gov Fights Poverty degree do you agree or disagree with this GDP per Capita Index Tolerance statement?” -0.1 0.0 0.1 0.2 95% C.I. In order to test the relationship between Source: 2008 AmericasBarometer by LAPOP, 21 countries, 27,546 respondents religious affiliations and support for democracy in 21 Latin American countries, I use 14 control In Figure 2, I consider two religious categories, variables at the individual level and one control Catholics as in the last model and, additionally, LDS and Jehovah’s Witnesses combined. In this 15 I want to acknowledge the important observations on this case, the rest of the denominations and non‐ subject about Evangelicals contributed by Luis E. Soto religious people represent the reference (USAID, Dominican Republic). © 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 5 of 13 category. The results show that both religious Figure 3. categories reduce support for democracy Support for Democracy in the Americas, 2008 compared to the remaining denominations and (Catholic, LDS and Non‐Christian) the non‐religious. Support for Democracy 2008 Eastern and Trad Rel LDS and Witness Catholic Figure 2. Church Attendance Rural to Urban Wealth Index Support for Democracy in the Americas, 2008 Female Percep Nat Eco Situation (Catholic, LDS and Non‐Catholic/Non‐LDS) Percep of Security Political Interest Education Level Interpersonal Trust Support for Democracy 2008 Political Knowledge Index Age LDS and Witness Trust Cath Church Catholic Gov Fights Poverty Church Attendance GDP per Capita Index Rural to Urban Tolerance Wealth Index -0.1 0.0 0.1 0.2 Female Percep Nat Eco Situation 95% C.I. Percep of Security Source: 2008 AmericasBarometer by LAPOP, 21 countries, 27,546 respondents Political Interest Education Level Interpersonal Trust Political Knowledge Index Age Trust Cath Church Figure 4 includes one additional category, Gov Fights Poverty GDP per Capita Index Tolerance Mainline Protestants. In this case, Evangelicals -0.1 0.0 0.1 0.2 and non‐ religious people represent the omitted 95% C.I. category. In this model there is no statistically Source: 2008 AmericasBarometer by LAPOP, 21 countries, 27,546 respondents significant effect for the new grouping, though this gradual inclusion of religious categories Figure 3 shows the effect of three religious reveals that control variables at the individual groups on support for democracy. In addition to and country level remain stable in the model. Catholics and LDS and Jehovah Witnesses, I now include Eastern and Traditional or Native Figure 4. religions. In this case, the reference category is Support for Democracy in the Americas, 2008 other Christian affiliations and people who do (Catholic, LDS, Non‐Christian and Protestant) not profess any religion. The main conclusion from Figure 3 is that non‐Christian affiliations Support for Democracy 2008 Protestant Eastern and Trad Rel are statistically unrelated to support for LDS and Witness Catholic democracy. Church Attendance Rural to Urban Wealth Index Female Percep Nat Eco Situation Percep of Security Political Interest Education Level Interpersonal Trust Political Knowledge Index Age Trust Cath Church Gov Fights Poverty GDP per Capita Index Tolerance

-0.1 0.0 0.1 0.2 95% C.I. Source: 2008 AmericasBarometer by LAPOP, 21 countries, 27,546 respondents

Finally, Figure 5 shows the final model, which includes five religious groups, where citizens who report no religious identification represent the reference category. In this case, the Protestant variable is now statistically

© 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 6 of 13 significant, as is the newly added Evangelical categorize specifically religious affiliations in the variable. Americas.17

A comparison between Figure 1 and the final A New Classification Scheme model in Figure 5 permits us to conclude that the positive relationship between religious Table 2 (see the appendix to this report) presents affiliations and support for democracy is driven the denominations to be used in the 2010 by two specific non‐Catholics groups: Mainline AmericasBarometer and the current affiliations Protestants and Evangelicals (compared to the considered by the WVS and the Latinobarómetro, reference category, those without any as points of comparison. affiliation). The new comprehensive list captures not only new and/or emerging religious groups, but also Figure 5. includes the general classifications presented by Support for Democracy in the Americas, 2008 (Religious Category Omitted: No Religion) the other two multinational projects. This formulation represents a new effort to improve Support for Democracy 2008 prior classifications by the AmericasBarometer.

Evangelical Protestant Eastern and Trad Rel LDS and Witness Specifically, the answers for the religious Attendance denomination question (Q3 in the Rural to Urban Wealth Index Female AmericasBarometer questionnaire) will include Percep Nat Eco Situation Percep of Security Political Interest a comprehensive list of religious affiliations Education Level Interpersonal Trust categorized as follows: Political Knowledge Index Age Trust Cath Church Gov Fights Poverty GDP per Capita Index Q3. What is your religion, if any? [Do not read Tolerance -0.1 0.0 0.1 0.2 options] 95% C.I. (1) Catholic Source: 2008 AmericasBarometer by LAPOP, 21 countries, 27,546 respondents (2) Mainline Protestant or Protestant non‐Evangelical (Protestant; Christian; Calvinist; Lutheran; Methodist;

Presbyterian; Disciple of Christ; Anglican; It is clear from this evidence that it is important Episcopalian; Moravian) to classify religious affiliations accurately and (3) Non‐Christian Eastern Religions (Islam; Buddhist; 16 precisely in the AmericasBarometer countries Hinduism; Taoist; Confucianism; Baha’i) as well as in other surveys because this (4) None (None; believes in a Supreme Entity but distinction permits us to test specific hypotheses does not belong to any religion) related to particular religious categories and (5) Evangelical and Pentecostal (Evangelical; sub‐groups. Existing theory and evidence Pentecostals; Church of God; Assemblies of God; available at the country level (Camp 1997; Gill Universal Church of the Kingdom of God; International Church of the Foursquare Gospel; Christ 2002; Hallum 2002) suggests that religious sub‐ Pentecostal Church; Christian Congregation; groups other than these six main categories may Mennonite; Brethren; Christian Reformed Church; be relevant politically or economically speaking. Charismatic non‐Catholic; Light of World; Baptist; For that reason, in order to have any means of Nazarene; Salvation Army; Adventist; Seventh‐Day testing this type of hypotheses, it is important to Adventist; Sara Nossa Terra) (6) LDS

16 Those countries are México, Guatemala, El Salvador, Honduras, Nicaragua, Costa Rica, Panamá, Colombia, Ecuador, Bolivia, Perú, Paraguay, Chile, Uruguay, Brazil, 17 I thank Prof. Elizabeth Zechmeister’s observation about Venezuela, Argentina, Dominican Republic, Haiti, Jamaica this distinction between religious categories and religious and Belize. sub‐groups. © 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 7 of 13 (7) Traditional Religions or Native Religions the Catholic Church leaders in relation to (Candomblé, Voodoo, Rastafarian, Mayan Traditional democracy (Gill 1998); and the capacity of the Religion; Umbanda; Maria Lonza; Inti; Kardecista, Catholic Church to mobilize citizens in politics Santo Daime, Esoterica) (Hagopian 2008) among other topics. In sum, (10) Does not believe in God (Agnostic; atheist) with the increased quality and precision of data (11) Jewish (Orthodox; Conservative; Reform) (12) Jehovah’s Witness concerning religious affiliations in the 2010 (88) DK/DR AmericasBarometer, we will gain a more detailed understanding of this aspect of the The main advantage of this approach is that the region in general and its relationships to critical AmericasBarometer will offer two variables in variables related to democracy and other facets its 2010 dataset derived from the religious of economic and political change and denomination question: a) the aforementioned development.20 categorization; and b) the detail of each religious affiliation.18 For example, the dataset will References contain the category Mainline Protestant and the database will detail whether or not the Banfield, Edward, C. 1958. The Moral Basis of a respondent belongs to the Moravian Church, or Backward Society. New York, NY: Free the Lutheran, Methodist or Presbyterian Church. Press. Barro, R. J., and R. McCleary. 2003. “Religion Concluding Remarks and Economic Growth across Countries.” American Sociological Review

68(5): 760‐781. This paper has attempted to provide an Bastian, J.P. 1993. “The Metamorphosis of Latin interpretation and critical understanding of American Protestant Groups. A Socio‐ religious affiliation’s political and economic historical Perspective.” Latin American significance and also offer a new, more thorough Research Review 28(2): 33‐61. classification that will allow a more complete Camp, Roderic Ai. 1997. Crossing Swords. Politics picture of religious denominations in the region. and . New York, NY: The resulting data can be used by scholars to Oxford University Press. test specific hypotheses related to particular Campbell, D. 2006. “Religious Threat in religious groups. Contemporary Presidential Elections.”

Journal of Politics 68(1): 104‐115. For example, the 2010 classification will allow Collins, Randall. 1997. “An Asian Route to analysis of the distinct relationship between LDS Capitalism: Religious Economy and the and Jehovah’s Witnesses affiliates, on the one origins of Self‐Transforming Growth in hand, and support for democracy, on the other, Japan.” American Sociological Review or support for specific public policies.19 62(6): 843‐865. Furthermore, the data will allow the study of Cruz, José Miguel. 2009. “Social Capital in the specific interactions among Catholics, Americas: Participation in Religious Protestants and Evangelicals in terms of vote Groups.” AmericasBarometer Insights choice, as the “religious threat” in the US Series. 15: 1‐5. (Campbell 2006); the effect of Evangelicals on Fontus, Fritz. 2001. Les Eglises Protestantes en Haïti. Communication et Inculturation. 18 This by‐product for the 2010 round will be possible thanks Paris, France: L’Harmattan. to the technological improvement based on handheld computers that will gather the information collected by the interviewers in many countries. I wish to thank Dominique Zephyr for calling my attention to this fact. 20 I also wish to thank Prof. Mitchell A. Seligson for his 19 I thank Matt Layton’s observation about this point. encourage and support in this project. © 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 8 of 13 Gill, Anthony. 1998. Rendering unto Caesar: The Commitment from 1980 to 1994.” Public Catholic Church and the State in Latin Opinion Quarterly 61(2): 288‐316. America. Chicago: University of Chicago Layman, Geoffrey. 2001. The Great Divide: Press. Religious and Cultural Conflict in ______. 2001. ʺReligion and Comparative American Party Politics. New York, NY: Politics.ʺ Annual Review of Political Columbia University Press. Science 4: 117‐138. Lenski, Gerhard. 1963. The Religious Factor. A ______. 2002. “Religion and Democracy in Sociological Study of Religionʹs Impact on South America. Challenges and Politics, Economics, and Family Life. Opportunities.” in Jelen and Wilcox. Garden City: Anchor Books. Religion and Politics in Comparative Levine, Daniel H. 1986. Religion and Political Perspective. The One, the Few, and the Conflict in Latin America. Chapel Hill, Many. New York, NY: Cambridge NC: The University of North Carolina University Press. Press. Hagopian, Frances. 2008. ʺLatin American Lewis, Oscar. 1966. “The Culture of Poverty.” Catholicism in an Age of Religious and Scientific American 215(4): 19‐25. Political Pluralism: A Framework for Lies, William. 2006. “The Chilean Church: Analysis.ʺ Comparative Politics 40(2): 149‐ Declining Hegemony?” in Manuel, 168. Wilcox, and Reardon. The Catholic Herberg, Will. 1955. Protestant, Catholic, Jew. An Church and the Nation‐State: Comparative Essay in American Religious Sociology. Perspectives. Washington, DC: Chicago, IL: The University of Chicago Georgetown University Press. Press. Mainwaring Scott. 1986. The Catholic Church and Huntington, Samuel. 1989. The Third Wave. Politics in Brazil, 1916‐1985. Stanford, Democratization in the Late Twentieth CA: Stanford University Press. Century. Norman: University of Martin, David. 1990. Tongues of Fire. The Oklahoma Press. Explosion of Protestantism in Latin Hallum, Anne M. 2002. “Looking for Hope in America. Cambridge, MA: Blackwell. Central America: The Pentecostal Marx, Karl. 1978. “On the Jewish Question.” in Movement.” in Jelen and Wilcox. Tucker, E. (ed) The Marx‐Engels Reader. Religion and Politics in Comparative New York, NY: W.W. Norton & Co. Perspective. The One, the Few, and the Magaloni, B. and A. Moreno. 2003. “Catching Many. New York, NY: Cambridge All Souls: the PAN and the Politics of University Press. Religion in Mexico.ʺ in Mainwaring and Harrison, Lawrence, E. 2000. Underdevelopment Scully. Christian Democracy in Latin as a State of Mind: The Latin American America: Electoral Competition and Regime Case. New York, NY: Cooper Square Conflicts. Stanford, CA: Stanford Press. University Press. Inglehart, Ronald. 1997. Modernization and Mishler, William and Richard Rose. 1999. “Five Postmodernization, Cultural, Economic and Years after the Fall: Trajectories of Political Change in 43 Societies. Princeton, Support for Democracy in Post‐ NJ: Princeton University Press. Communist Europe.” in Norris. Critical Kahn, Herman. 1979. World Economic Citizens: Global Support for Democratic Development: 1979 and Beyond. Governance. Oxford: Oxford University Indianapolis, IN: Hudson Institute. Press. Layman, Geoffrey. 1997. “Religion and Political Noland, Marcus. 2005. ”Religion and Economic Behavior in the United States: The Performance.” World Development 33(8): Impact of Beliefs, Affiliations, and 1215‐1232. © 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 9 of 13 Przeworski, A., J.A. Cheibub, and F. Limongi. 1998. “Culture and Democracy.” in World Culture Report: Culture, Creativity and Markets. Paris: UNESCO. Rose, Richard and Doh Chull Shin. 2001. “Democratization Backwards: The Problem of Third‐Wave Democracies.” British Journal of Political Science 31(2): 331‐354. Sherman, Amy. 1997. The Soul of Development. Biblical and Economic Transformation in Guatemala. New York, NY: Oxford University Press. Stein, Andrew J. 2000. “Religion, Political Preferences and Protest Action in Central America: Nicaragua, El Salvador, and Guatemala.” in Burdick and Hewitt (ed.) The Church at the Grassroots in Latin America: Perspectives on Thirty Years of Activism. Westport, CT: Praeger Publishers. Stoll, David. 1993. Rethinking Protestantism in Latin America. Philadelphia: Temple University Press. Swank, Duane. 1996. “Culture, Institutions and Economic Growth: Theory, Recent Evidence and the Role of Communitarian Polities.” American Journal of Political Science 40(3): 660‐679. Wald, K.D., and C. Wilcox. 2006. “Getting Religions: Has Political Science Rediscovered the Faith Factor?” American Political Science Review 100(4): 523‐529. Weber, Max. 1958. The Protestant Ethic and the Spirit of Capitalism. New York, NY: Scribner. Turner, Frederick C. 1971. Catholicism and Political Development in Latin America. Chapel Hill, NC: The University of North Carolina Press.

© 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 10 of 13 Table 1.

Religious Affiliations (%) by Latin American country, AmericasBarometer, round 2008

Country Catholic Mainline Protestant Non Christian No religion Evangelical LDS and JW Trad Religions Total of cases

Argentina 77.1% 0.9% 1.9% 15.9% 3.3% 0.6% 0.2% 1,486 Belize 46.6% 17.9% 0.2% 7.3% 7.2% 2.7% 18.1% 1,552 Bolivia 81.8% 2.6% 0.4% 3.3% 10.3% 1.7% 0.1% 3,003 Brazil 69.5% 8.0% 0.4% 6.4% 11.5% 1.1% 3.0% 1,497 Chile 68.3% 2.2% 0.3% 13.3% 13.6% 2.1% 0.2% 1,527 Colombia 82.7% 3.1% 0.4% 6.6% 6.6% 0.5% 0.0% 1,503 Costa Rica 69.0% 13.7% 1.1% 8.2% 7.1% 0.9% 0.1% 1,500 Dominican Republic 67.6% 8.1% 0.3% 10.2% 12.1% 1.6% 0.1% 1,507 Ecuador 83.8% 0.9% 0.2% 5.8% 7.4% 1.9% 0.0% 3,000 El Salvador 53.8% 9.6% 0.3% 12.4% 22.2% 1.6% 0.1% 1,549 Guatemala 56.5% 9.3% 0.3% 10.5% 22.0% 1.2% 0.2% 1,538 Haiti 55.5% 29.1% 0.2% 7.0% 3.2% 1.7% 3.2% 1,536 Honduras 68.5% 7.2% 1.3% 10.2% 12.3% 0.4% 0.0% 1,522 Jamaica 4.6% 32.5% 1.0% 7.8% 36.3% 13.3% 4.6% 1,499 Mexico 84.7% 1.2% 0.3% 7.7% 4.4% 1.6% 0.1% 1,560 Nicaragua 57.1% 7.6% 0.7% 12.5% 20.7% 1.2% 0.3% 1,540 Panama 79.3% 6.4% 1.1% 3.5% 8.5% 1.0% 0.1% 1,536 Paraguay 88.7% 2.4% 0.1% 1.8% 5.8% 1.2% 0.0% 1,166 Peru 80.0% 3.1% 0.3% 4.9% 9.4% 2.2% 0.1% 1,500 Uruguay 52.8% 1.1% 1.0% 34.6% 6.9% 1.9% 1.6% 1,500 Venezuela 82.5% 1.1% 0.6% 9.3% 4.7% 1.5% 0.3% 1,500 Total 68.4% 7.5% 0.6% 9.1% 11.1% 2.0% 1.4% 34,521

Mainline Protestant or Protestant non-Evangelical = Adventist, Baptist, Calvinist, The Salvation Army, Lutheran, Methodist, Nazarene, Presbyterian. Non-Christian Religions = Jewish, Muslims, Buddhists, Hinduisms, Taoists. Evangelical and Pentecostal = Pentecostals, Charismatic non-Catholics, Light of World. LDS and JW = Latter-Day Saints, Jehovah’s Witness, Spiritualists and Seventh-Day Adventists. Traditional Religions or Native Religions = Candomblé, Voodoo, Rastafarian, Mayan Traditional Religion. None = secularist, atheist, do not believe in God. Valid cases = 34,223 respondents. Missing cases = 298 (0.9%)

© 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 11 of 13 Table 2

Comprehensive list of Religious Affiliations in Latin America, by survey project

AmericasBarometer 2010 (project) WVS (1990‐2005) LB (1995‐2007)

Catholic Catholic Roman Catholic Catholic

Catholic (does not follow rules)

Orthodox Orthodox

Christian Christian Christian (in general)

Mainline Protestant Anglican

Calvinist

Disciple of Christ

Episcopalian

Lutheran

Methodist Evangelical (Methodist)

Moravian

Presbyterian

Protestant Protestant (in general) Protestant (in general)

Evangelical Adventist Adventist

Baptist Evangelical (Baptist)

Charismatic non‐Catholic

Assemblies of God

Church of God

Universal Church of the Kingdom of God

International Church of the Foursquare Gospel

Light of World

Nazarene

Pentecostals Pentecostal Evangelical (Pentecostal)

Salvation Army

Brethren

Christ Pentecostal Church

Christian Congregation

© 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 12 of 13 AmericasBarometer 2010 (project) WVS (1990‐2005) LB (1995‐2007)

Evangelical (cont) Mennonite

Christian Reformed Church

Seventh‐Day Adventist Seven‐day Adventist

Evangelical Evangelical (in general) Evangelical (no specific denomination)

Jehovah’s Witness Jehovah’s Witness Jehovah Witnesses Jehovah Witness

LDS Latter‐Day Saints (LDS) LDS

Non‐Christian Buddhist Buddhist

Eastern Religions Bahaʹi

Confucianism

Hinduism Hindu

Islam Islam

Taoist

Judaism Orthodox, Conservative, Reform Jew Jewish

Traditional Candomblé Candomblé

Religions Voodoo

Rastafarian

Mayan Traditional Religion

Maria Lonza

Inti

Kardecista Spiritism/Esoterism/Occultism Spiritism/Kardecism

Umbanda Umbanda Umbanda

Santo Daime

Esoterica

None None None None

Believes in a Supreme Entity but does not belong to any religion Believer (does not belong to any church)

Atheist Atheist Atheist

Agnostic Agnostic

Blank spaces = Not asked

© 2009, Latin American Public Opinion Project, “Insights” series www.AmericasBarometer.org Page 13 of 13